[00:07:54] --- sxw has left [07:02:28] --- Russ has become available [07:21:21] --- deason has become available [07:31:58] --- Derrick Brashear has left [07:49:53] --- Russ has left: Disconnected [08:04:19] --- Derrick Brashear has become available [08:24:49] --- Russ has become available [09:04:48] --- dev-zero@jabber.org has become available [09:04:48] --- dev-zero@jabber.org has left: offline [09:16:54] --- cclausen has left [10:18:19] --- reuteras has left [11:10:44] --- mmeffie has become available [12:50:19] --- dev-zero@jabber.org has become available [12:50:23] --- dev-zero@jabber.org has left: offline [13:01:17] --- mmeffie has left [13:11:35] --- Russ has left: Disconnected [13:37:34] --- Russ has become available [13:44:58] --- Simon Wilkinson has become available [13:55:29] Hey Simon. [13:55:50] Hi Russ [14:03:57] source for the cliff stoll cookies: http://www.sinc.sunysb.edu/Staff/bronwen/chocolate.html [14:14:54] opening the windows was a good idea. it was getting a bit warm in here. [14:15:12] Indeed. Much nicer now, as long as they don't start singing again. [14:15:15] the windows only got closed because of people making noises outside [14:15:26] I was sort enjoying the people changing [14:15:28] chanting* [14:16:04] Oi! You're suppose to be paying attention! [14:29:05] /tmp/buildd/libafs-perl-2.6.1/src/../blib/arch/auto/AFS/AFS.so: undefined symbol: UV_PartitionInfo [14:31:55] UV_PartitionInfo64 [14:32:34] i guess i can give you the backwards compat function as a patch if you want it; UV_* isn't really an API [14:34:40] Did something similar happen to rx_NewCell? [14:34:48] Er, rx_NewCall. [14:35:04] * Russ can patch it rather than adding backward compat to OpenAFS -- that's okay. [14:35:07] Thanks for the new function name. [14:35:10] * Russ starts hacking. [14:35:18] NewCall? no? [14:44:12] --- stevenjenkins has left [14:44:38] --- deason has left: Lost connection [14:46:25] --- deason has become available [14:53:43] ultraviolet? [15:03:04] Russ: We have patches for AFS perl. [15:03:20] Let me dig them out, it will probably save you some pain. [15:04:10] That would be great. [15:04:17] I think I may have everything, but not sure. [15:05:57] /afs/inf.ed.ac.uk/user/s/sxw/Public/perl-AFS-*.diff [15:08:31] Huh, those are completely orthogonal to the ones I have. [15:09:32] And I messed up the PartInfo change, in that it's apparently more complex than I thought. [15:09:58] Oh, wait, no, I just didn't see all the files. [15:10:45] You probably don't need -extrafunctions or -devel-15. Extrafunctions adds support for bos setcellname, adduser, removeuser, addhost, removehost and setrestart [15:11:31] --- stevenjenkins has become available [15:11:42] Yeah, the key one was the partinfo one. [15:12:35] Had you already sent this stuff to Norbert, out of curiosity? [15:13:10] I dunno, actually. I didn't write it - the guy who's integrating AFS with our configuration management system wrote it all. I think he was going to, but I'll check. [15:13:29] We so need a real library API. [15:13:36] --- jhutz@jis.mit.edu/owl has become available [15:13:40] * Russ wants the AFS Perl module on 64-bit systems. [15:14:03] Uh, I think I got that to build with only minor issues. [15:14:05] Use too. [15:14:08] Us too, even. [15:14:22] I got it to work a while back, but I built the whole of it with -fPIC. [15:14:28] The whole of AFS, even. [15:14:29] jhutz: You have to build all of OpenAFS with -fPIC. [15:14:35] Which we can do, but ugh. [15:15:07] Our plan is to implement all of the fs commands we care about by using the magic ioctl stuff. [15:15:24] That way we can configure 64 bit AFS clients. And we don't really care about 64bit AFS servers at the moment. [15:15:33] Which means you're basically inventing a new AFS client library. [15:16:24] Yeh. Which sucks, but it's nicer than any of the alternatives I've seen so far. [15:16:27] I suspect the result may actually be interesting for OpenAFS itself if it has a well-thought-out API and a clear design and documentation. Although rewrites aren't necessarily the right way to solve problems. [15:16:50] But there's so much interlocking cruft in the current AFS libraries that it's horribly intimidating to figure out where to start. [15:17:31] Although that only gives you the stuff that can be done via ioctls, not the protocol stuff that requires Rx, I assume. [15:17:47] Basically, I have a co worker who's building all of this stuff. He definitely doesn't have time to work out how to solve OpenAFS's library problem, but he does have time to write enough XS to get the ioctls he needs working. [15:18:01] Yes. Essentially, we're just looking at configuring the cache manager with it, so it's all local. [15:18:05] Right. [15:18:13] Which is an interesting problem, but not really the big one. [15:18:24] Indeed. [15:29:14] --- deason has left [15:30:22] --- deason has become available [16:53:39] --- kaduk@mit.edu/owl has become available [17:25:13] --- Derrick Brashear has left [17:25:51] --- Simon Wilkinson has left [17:38:45] --- Russ has left: Disconnected [18:20:05] --- cclausen has become available [21:04:12] --- Russ has become available [21:33:59] --- Russ has left: Replaced by new connection [21:33:59] --- Russ has become available [21:58:40] --- deason has left [22:18:17] --- cclausen has left [22:47:05] --- Russ has left: Disconnected [23:14:11] --- reuteras has become available [23:17:09] --- Simon Wilkinson has become available [23:19:07] --- Derrick Brashear has become available [23:37:22] --- Simon Wilkinson has left